Tiger deaths in Vietnam due to infection with H5N1 highly pathogenic avian influenza virus bearing mutations associated with mammalian host adaptation

Recently, infections with the H5N1 subtype of highly pathogenic avian influenza virus (H5N1-HPAIV) in mammals have been reported worldwide, including in cows in the United States and successive human cases in Cambodia. In Vietnam, 47 tigers and three leopards died from H5N1-HPAIV infection between August and October 2024. This study aimed to determine the origin of the H5N1 strains that infect tigers in Vietnam and to identify specific mutations associated with mammalian infection. Specimens were collected from tigers that died of suspected H5N1-HPAIV infection in southern Vietnam in September-October 2024. RNA was extracted and subjected to whole-genome sequencing. Time-stamped phylogenetic analysis was performed using H5N1 sequences recently detected in Vietnam and neighbouring countries. Phylogenetic results revealed that the strain found in tigers (Tiger H5N1 strain) belonged to clade 2.3.2.1e and has been genetically close to the H5N1-HPAIV lineage responsible for ongoing human infections in Cambodia since 2023. Tiger H5N1 strains harboured several amino acid substitutions associated with mammalian host adaptation or transmissibility, such as E627K in polymerase basic protein 2, similar to the Cambodian human H5N1 strains. This mammalian-adapted H5N1 lineage should be continuously monitored in poultry and mammals, including humans, in Vietnam to prevent further transmission.

Main text

Avian influenza virus (AIV) primarily resides in waterfowl and can be transmitted to mammals, including dairy cows, cats, and humans [1], causing a substantial disease burden and economic losses. From October 2024 to May 2025, 2,694 highly pathogenic avian influenza (HPAI) outbreaks were reported in poultry and wild birds worldwide, including 12 mammalian transmissions [2]. Transmission typically occurs from birds to mammals; however, virus–host dynamics have shifted markedly in recent years. For instance, more than 19,000 sea lions reportedly succumbed to H5N1 highly pathogenic AIV (H5N1-HPAIV) infections in 2023 in Peru and Chile [3]. In Cambodia, continuous human H5N1-HPAIV infections have been reported since February 2023, with 27 cases and 12 deaths (mortality rate: 44.4%) as of July 5, 2025 [4].

In Vietnam, a 21-year-old university student died of H5N1-HPAIV infection in Khanh Hoa Province in March 2024 [5]. During poultry surveillance conducted in Vietnam from 2017 to 2022, an average of 3.6% of samples tested positive for the H5 subtype of AIV. In 2022 alone, this rate rose to 5.8%, indicating an increase in poultry infections in recent years [6]. Clade 2.3.2.1e of H5N1-HPAIV has predominantly circulated in southern Vietnam, whereas clade 2.3.4.4b strains have been prevalent nationwide since 2021 [7]. In October 2024, it was reported that at least 50 mammals (47 tigers and three leopards) had died of H5N1-HPAIV infection in Vietnam [8]. Captive mammals are rarely infected with H5N1-HPAIV, and when infected, typically die within a short period. This study aimed to determine the origin of the H5N1 strains that infected tigers in Vietnam (Tiger H5N1 strains) through phylogenetic analysis and to identify any mutations associated with mass mammalian mortality.

Intestinal samples were collected from tigers that died of suspected H5N1-HPAIV infection in the Long An and Dong Nai provinces of Vietnam in September – October 2024 (Supplementary Figure S1). The tiger samples were transported to the National Institute of Hygiene and Epidemiology and National Center for Veterinary Diagnosis and whole-genome sequencing was performed. Time-stamped and maximum-likelihood trees were inferred using complete open reading frame (ORF) sequences. Detailed methods are available in the Supplementary Material.

BLAST analysis of the consensus sequences obtained in this study revealed high homology to previous H5N1 strains detected in Vietnam and neighbouring countries, such as Laos and Cambodia (Supplementary Table S1). Accordingly, time-stamped phylogenetic trees were constructed for the haemagglutinin (HA) and polymerase basic protein 2 (PB2) genes using reference sequences from neighbouring countries to clarify the origin of the H5N1 strain detected in tigers in Vietnam (Tiger H5N1 strain). These strains belonged to clade 2.3.2.1e and were genetically similar to the H5N1-HPAIV lineage, which has caused ongoing human infections in Cambodia since 2023 (Figure 1A). It was inferred that the H5N1-HPAIV lineage causing human infections in Cambodia (Cambodian human H5N1 strain) entered Vietnam at the end of 2023 and spread among poultry in the first half of 2024, subsequently infecting tigers from August to October 2024 (Figure 1A). To validate the time-stamped analysis, a phylogenetic tree was constructed using the HA ORF sequences of all H5N1 strains detected in Vietnam, Cambodia, and Laos from 2020 to the present, and the topology of both phylogenetic trees was found to be consistent (Supplementary Figure S2). Similar results were confirmed for PB2 and all other segments compared to the HA segment (Figure 1B, Supplementary Figure S3A–S3F). Cambodian human H5N1 strains harboured several amino acid substitutions associated with mammalian host adaptation or transmissibility in mammals, such as N158D and T160A in HA, and E627 K in PB2, which persisted in strains that infected poultry and tigers in Vietnam (Figure 1A) [9,10]. HA proteins of the Tiger and Cambodian human H5N1 strains contained PQRERRRKR↓GLF with multiple basic amino acids at the cleavage site, indicating high pathogenicity to poultry and mammals. The neuraminidase (NA) protein of the Tiger H5N1 strain possessed no amino acid substitutions associated with NA inhibitor resistance (Supplementary Table S2) [11]. In contrast, the Tiger H5N1 strain harboured unique amino acid substitutions at positions 188, 189, and 242 in HA and 555 in PB2 (Figure 1A and 1B, Supplementary Table S2). These findings suggest that the Tiger H5N1 strains originally harboured mammalian-adaptive amino acid substitutions and that the lineage evolved uniquely during the outbreak among domestic poultry and tigers in Vietnam.

Figure 1.

Phylogenetic trees of haemagglutinin (HA) and polymerase basic protein 2 (PB2) open reading frame (ORF) sequences of Tiger H5N1 avian influenza virus (AIV). Time-stamped Bayesian trees of the (A) HA and (B) PB2 ORF were constructed using sequences of H5N1 strains detected in Vietnam and neighbouring countries. Amino acid substitutions associated with mammalian adaptation and transmission are shown to the right of each strain. Colours denote H5N1 lineages: red, Tiger H5N1 strains; blue, past Vietnamese H5N1 strains; green, Cambodian H5N1 human strains. The corresponding H5N1 clades are shown on the right. Numbers and dates in the strain name indicate the GISAID accession numbers and sample collection dates, respectively.

Discussion

In this study, the Tiger H5N1 viruses were classified as clade 2.3.2.1e and were found to be closely related to the lineage that has caused human infections in Cambodia in recent years. In fact, the Tiger H5N1 strains and human-associated Cambodian strains possessed amino acid substitutions related to mammalian adaptation (E627 K in PB2) and transmission (N158D and T160A in HA), which are characteristic of mammalian infections. Human infections with H5N1 clade 2.3.2.1e viruses in Cambodia showed a high mortality rate of 44.4% (12 deaths in 27 cases) [4], whereas a total of 70 human infections with H5N1 clade 2.3.4.4b strains were reported in the US during the period spanning from 2024 to mid-February 2025 [12]. Of these, only a single case of fatality was observed in Louisiana in January 2025. There have been no subsequent reports of cases in the US since March 2025 [13]. The mammalian-adapted H5N1 clade 2.3.2.1e lineage may cause human infections in Vietnam soon, as it has been spreading in southern Vietnam and should be continuously monitored to protect public health.

Phylogenetic analysis of the HA gene of all strains detected in Vietnam and neighbouring countries from 2020 to the present showed that most H5N1 strains prevalent in Vietnam in recent years belong to clade 2.3.4.4b, whereas only a small number of clade 2.3.2.1e strains were identified, including the Tiger H5N1 strains, possibly introduced from Cambodia or Laos (Supplementary Figure S2). H5N1-HPAIV of clade 2.3.4.4b also remains prevalent in poultry in Vietnam. This clade has raised significant concerns due to its widespread infection in dairy cattle and livestock workers in the United States and should be closely monitored, although it differs from the Tiger H5N1 strains.

The most recent report investigating the lineage and distribution of the H5N1 influenza virus in Vietnam was published in 2019 [14]. This report indicated that clade 2.3.2.1 was distributed in central and southern Vietnam, while clade 2.3.4.4 was predominantly found in northern Vietnam [14]. However, since 2020, clade 2.3.4.4b has rapidly spread worldwide, resulting in avian influenza outbreaks reported in 84 countries between 2022 and 2023 [15]. In Vietnam, clade 2.3.4.4b has also expanded nationwide since 2020, and as of 2025, clade 2.3.4.4b has become the majority within the country (Supplementary Figure S2). At present, clades 2.3.2.1e and 2.3.4.4b are found in the same geographical area in southern Vietnam. Notably, the identification of clade 2.3.2.1e viruses with mammalian adaptation potential in this study has given rise to concerns regarding the acquisition of mammalian adaptation capacity through genetic reassortment. Acquisition of mammalian adaptation capacity by a highly infectious lineage such as clade 2.3.4.4b could potentially trigger the next pandemic. Consequently, it should be essential to carefully monitor the emergence of reassortants in Vietnam over the forthcoming decade.

Funding Statement

This work was supported by the Japan Agency for Medical Research and Development (AMED) under grant number JP24wm0125006 and SCARDA under grant number JP223fa627004.

Author contributions

MA and HA drafted the manuscript. MA, NTN, NDT, NTD, DTV and HA performed the experiments and analysed the data. NDT, NTD, DTV, FH, NLKH, and LTQM contributed to the sample collection. LTQM and HA designed and organized the study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethics approval

The study was approved by the Institutional Review Board of National Institute of Hygiene and Epidemiology, Vietnam (IRB-VN01057-20/2015), and Research Ethics Committee of Nagasaki University Institute of Tropical Medicine, Japan (240919318).

Data availability statement

The sequencing data generated in this study were deposited in GISAID under the accession numbers EPI_ISL_19520619, EPI_ISL_19520621, EPI_ISL_19520622, EPI_ISL_19520627, EPI_ISL_19520629, EPI_ISL_19520689, and EPI_ISL_20081455.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/22221751.2025.2582252.